File Organizations and Indexing

R&G Chapter 8"If you don't find it in the index, look very carefully through the entire catalogue."

-- Sears, Roebuck, and Co., Consumer's Guide, 1897

Context

Query Optimizationand Execution

Relational Operators

Files and Access Methods

Buffer Management

Disk Space Management

DB

Alternative File Organizations

Many alternatives exist, each good for some situations, and not so good in others:– Heap files: Suitable when typical

access is a file scan retrieving all records.

– Sorted Files: Best for retrieval in search key order, or only a `range’ of records is needed.

– Clustered Files (with Indexes): Coming soon…

Cost Model for Analysis

We ignore CPU costs, for simplicity:– B: The number of data blocks– R: Number of records per block– D: (Average) time to read or write disk block– Measuring number of block I/O’s ignores

gains of pre-fetching and sequential access; thus, even I/O cost is only loosely approximated.

– Average-case analysis; based on several simplistic assumptions.

Good enough to show the overall trends!

Some Assumptions in the Analysis

• Single record insert and delete.• Equality selection - exactly one

match (what if more or less???).• Heap Files:

– Insert always appends to end of file.

• Sorted Files:– Files compacted after deletions.– Selections on search key.

Cost of Operations

B: The number of data pagesR: Number of records per pageD: (Average) time to read or write disk page

Heap File Sorted File Clustered File

Scan all records

Equality Search

Range Search

Insert

Delete

Cost of Operations

B: The number of data pagesR: Number of records per pageD: (Average) time to read or write disk page

Heap File Sorted File Clustered File

Scan all records

BD BD

Equality Search

Range Search

Insert

Delete

Cost of Operations

B: The number of data pagesR: Number of records per pageD: (Average) time to read or write disk page

Heap File Sorted File Clustered File

Scan all records

BD BD

Equality Search

0.5 BD (log2 B) * D

Range Search

Insert

Delete

Cost of Operations

B: The number of data pagesR: Number of records per pageD: (Average) time to read or write disk page

Heap File Sorted File Clustered File

Scan all records

BD BD

Equality Search

0.5 BD (log2 B) * D

Range Search

BD [(log2 B) + #match pg]*D

Insert

Delete

Cost of Operations

B: The number of data pagesR: Number of records per pageD: (Average) time to read or write disk page

Heap File Sorted File Clustered File

Scan all records

BD BD

Equality Search

0.5 BD (log2 B) * D

Range Search

BD [(log2 B) + #match pg]*D

Insert 2D ((log2B)+B)D (because R,W 0.5)

Delete

Cost of Operations

B: The number of data pagesR: Number of records per pageD: (Average) time to read or write disk page

Heap File Sorted File Clustered File

Scan all records

BD BD

Equality Search

0.5 BD (log2 B) * D

Range Search

BD [(log2 B) + #match pg]*D

Insert 2D ((log2B)+B)D

Delete 0.5BD + D ((log2B)+B)D (because R,W 0.5)

Indexes

• Sometimes, we want to retrieve records by specifying the values in one or more fields, e.g.,– Find all students in the “CS” department– Find all students with a gpa > 3

• An index on a file is a disk-based data structure that speeds up selections on the search key fields for the index.– Any subset of the fields of a relation can be

the search key for an index on the relation.– Search key is not the same as key (e.g.

doesn’t have to be unique ID).• Can have multiple (different) indexes per file.

– E.g. sort by age, with an index on salary and name.

Index Breakdown

• Index Data Structure– Tree-based, hash-based, other– What can the index speed up, and how much?

• Associating index entries with records– primary vs. secondary indexes, handling

duplicates– what kind of info is the index actually storing?

• Clustered vs. Unclustered Indexes• Single Part vs. Multi-Part Keys

– Multi-part key = “Composite Indexes”

Data structures

• What kinds of selections do they support?– Selections of form field <op> constant– Equality selections (op is =)– Range selections (op is one of <, >, <=, >=,

BETWEEN)

• Hash-based structures (how to grow/shrink)– Key problem on disk is handling growth– Extendible/Linear Hashing (Chap 11)

• Tree based structures– Why not binary tree? Estimate log2(1M) * D

– B-Tree, B+-Tree (Chap 10)

Wide World of Index Structures

• 2-dimensional ranges (“east of Berkeley and west of Truckee and North of Fresno and South of Eureka”)– Or distances (“within 2 miles of Soda Hall”)– Or n-dimensional – One common n-dimensional index: R-tree

• Supported in Oracle and Informix– See http://gist.cs.berkeley.edu for research on this topic

• Nearest neighbor (“closest BMW dealer”)• Ranking queries (“10 best Berkeley Thai restaurants on

price and atmosphere”)– these are hard to support!

• Regular expression matches– Suffix Trees

• XML path matches– DataGuide, 1-Index

Primary vs. Secondary Index

• Primary index search key must contain a “real” key, usually primary key– e.g., social security #, ISBN, etc.– No duplicate support– Store record in the index?

• Secondary index– e.g., “eye color”, “year of birth”, etc.– Duplicate support required– Use RID or primary key to refer to record?

Alternatives for Data Entry k* in Index

• Three alternatives: Actual data record (primary index

only) <k, rid of matching data record> <k, list of rids of matching data

records>

• Choice is orthogonal to the indexing technique.

Alternatives for Data Entries (Contd.)

• Alternative 1: Actual data record – Use index structure as the file structure – Saves pointer lookups for primary index

searches– Adds a primary index lookup for secondary

index access!Index nodes have all the issues of record management

Alternatives for Data Entries (Contd.)

Alternative 2

<k, rid of matching data record>

and Alternative 3

<k, list of rids of matching data records>– If heap file is used (no alt 1 indexes), then physical rid

can be used instead of primary key to refer to records– Alternative 3 more compact than Alternative 2, but

leads to variable sized data entries even if search keys are of fixed length.

– Even worse, for large rid lists the data entry would have to span multiple blocks! (how many?)

– Typical solution: add primary key or rid to end of secondary keys, and use Alternative 2!

Index Classification

• Clustered vs. unclustered: If order of data records is the same as, or `close to’, order of index data entries, then called clustered index.– A file can be clustered on at most one search key.– Cost of range scans through index varies greatly

based on whether index is clustered or not!– “Alternative 1-light”

• Alternative 1 implies clustered, but not vice-versa.• Use Physical RID in secondary index (why is this good?)

Clustered vs. Unclustered Index

• Suppose that Alternative (2) is used for data entries, and that the data records are stored in a Heap file.– To build clustered index, first sort the Heap file (with some

free space on each block for future inserts). – Overflow blocks may be needed for inserts. (Thus, order of

data recs is `close to’, but not identical to, the sort order.)

Index entries

Data entries

direct search for

(Index File)

(Data file)

Data Records

data entries

Data entries

Data Records

CLUSTERED UNCLUSTERED

Unclustered vs. Clustered Indexes

• What are the tradeoffs????• Clustered Pros

– Efficient for range searches– May be able to do some types of

compression– Possible locality benefits (related data?)– ???

• Clustered Cons– Expensive to maintain (on the fly or

sloppy with reorganization)– Pages tend to be only 2/3 full!

Cost of Operations

B: The number of data pagesR: Number of records per pageD: (Average) time to read or write disk page

Heap File Sorted File Clustered File

Scan all records

BD BD 1.5 BD

Equality Search

0.5 BD (log2 B) * D (logF 1.5B) * D

Range Search

BD [(log2 B) + #match pg]*D

[(logF 1.5B) + #match pg]*D

Insert 2D ((log2B)+B)D ((logF 1.5B)+1) * D

Delete 0.5BD + D ((log2B)+B)D (because R,W 0.5)

((logF 1.5B)+1) * D

Composite Search Keys

• Search on a combination of fields.– Equality query: Every field value

is equal to a constant value. E.g. wrt <age,sal> index:

• age=20 and sal =75– Range query: Some field value is

not a constant. E.g.:• age > 20; or age=20 and sal >

10

• Data entries in index sorted by search key to support range queries.– Lexicographic order – Like the dictionary, but on fields,

not letters!

sue 13 75

bob

cal

joe 12

10

20

8011

12

name age sal

<sal, age>

<age, sal> <age>

<sal>

12,20

12,10

11,80

13,75

20,12

10,12

75,13

80,11

11

12

12

13

10

20

75

80

Data recordssorted by name

Data entries in indexsorted by <sal,age>

Data entriessorted by <sal>

Examples of composite keyindexes using lexicographic order.

Summary• File Layer manages access to records in pages.

– Record and page formats depend on fixed vs. variable-length.

– Free space management an important issue.– Slotted page format supports variable length records

and allows records to move on page.

• Many alternative file organizations exist, each appropriate in some situation.

• If selection queries are frequent, sorting the file or building an index is important.– Hash-based indexes only good for equality search.– Sorted files and tree-based indexes best for range

search; also good for equality search. (Files rarely kept sorted in practice; B+ tree index is better.)

• Index is a collection of data entries plus a way to quickly find entries with given key values.

Summary (Contd.)

• Data entries in index can be actual data records, <key, rid> pairs, or <key, rid-list> pairs.– Choice orthogonal to indexing structure (i.e., tree, hash,

etc.).

• Usually have several indexes on a given file of data records, each with a different search key.

• Indexes can be classified as clustered vs. unclustered• Differences have important consequences for

utility/performance.• Catalog relations store information about relations,

indexes and views.